If $\sum x_n$ converges, is the sequence $a_k =\sum_{n=1}^k x_{2n}^2$ Cauchy?

I suspect that it isn't but a bit stuck on finding a counter example. For a valid counterexample, I believe I need a series that converges but the squared even partial sum does not.

Writing the definition of Cauchy, we know a sequence $(x_n)$ is Cauchy if for every $\epsilon > 0$, there exists $N \in \mathbb{N}$ such that $m, n \geq N \implies |x_m -x_n| < \epsilon$. But it seems easier to just use that convergence implies Cauchy rather than work directly from the definition. Any ideas/tips on what exactly I'm not seeing?

  • $\begingroup$ You cannot find a counter-example if the sequence is nonnegative since then $x_n^2 \leq x_n$ for all sufficiently large $n$. So... $\endgroup$ – Michael Oct 15 '18 at 1:35
  • $\begingroup$ If $x_n=(-1)^n/\sqrt n$ then $\sum_n x_n$ converges but $\sum_n x_n^2$ diverges. (Yes, I am aware that this isn't your exact problem.) $\endgroup$ – Lord Shark the Unknown Oct 15 '18 at 1:35
  • $\begingroup$ @Michael What if $x_n<0$? $\endgroup$ – Lord Shark the Unknown Oct 15 '18 at 1:37
  • $\begingroup$ @LordSharktheUnknown : I don't follow your comment directed to me ("@Michael what if..."), I think you meant something else or you misread my first comment. Your first comment seems in teh same spirit as mine but gives a bit of a larger hint. $\endgroup$ – Michael Oct 15 '18 at 1:38
  • $\begingroup$ I see now there is a third hinter. With all these "hints" the problem will soon be solved without the asker's involvement! $\endgroup$ – Michael Oct 15 '18 at 1:45

Hint: Let $\{y_n\}$ be a sequence of positive reals that converges to $0$, and take

$$x_n=\begin{cases} -y_{n/2} & \mathrm{if\ }n\equiv 0\bmod 2 \\ y_{(n+1)/2} & \mathrm{if\ }n\equiv 1\bmod 2,\end{cases}$$

or in other words $x_1,x_2,\cdots,$ is $y_1,-y_1,y_2,-y_2,y_3,-y_3\cdots$.

Then, your second series is

$$\sum_{n=1}^k y_n^2.$$

Can you find a sequence $\{y_n\}$ that converges to $0$ for which the above does not converge?

  • $\begingroup$ This doesn't converge to 0, but it seems this would work for the question, $\sum \frac{(-1)^{n+1}}{n}$? Converges but the above squared doesn't converge. $\endgroup$ – SS' Oct 15 '18 at 2:12
  • 1
    $\begingroup$ @SS' Isn't the square of it $\sum \frac{1}{n^2}$, which converges? However, that idea almost works - what if you put a slower-growing function in the denominator? $\endgroup$ – Carl Schildkraut Oct 15 '18 at 2:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.